Despite clear evidence of efficacy, little is known about the mechanisms of change underlying motivational interviewing. This project will evaluate a promising theoretical model for the effectiveness of motivational interviewing (MI) by examining client and therapist in-session behaviors during Project MATCH sessions. Our causal model is that Ml is successful because (and to the extent that) the counselor uses specific therapeutic behaviors that elicit client speech favoring change. If this model is supported, future motivational interviewers will be better able to focus on the components of this method that influence outcome, reducing irrelevant superstitious therapeutic behaviors. In addition, illuminating a relationship between in-session client behaviors and drinking outcomes would have the advantage of providing therapists with ongoing immediate feedback about the effectiveness of their clinical interventions, whatever their theoretical orientation. This study offers the advantage of testing theory-driven hypotheses about casual mechanisms for Ml without the expense of a new clinical trial. The integrity of the Project MATCH motivational enhancement therapy (MET) and the comprehensive client follow up data already collected in that project provide a vehicle for cost-effective investigation of mechanisms of action in MI.